If you're sitting across from a demand letter from a patent troll, the first thing to know is this: you are not the only target. You're target number 47 in a campaign that started six months ago and will continue against another 100 companies after yours. The plaintiff is running a process — and the process is what you're really up against, not the patent.
This is the step-by-step playbook of a typical patent-troll campaign. Once you see the volume-based extortion scheme for what it is, you stop fighting the wrong opponent.
The shape of the cycle
flowchart TD
A["1. Acquire patent<br/>$10k–$200k"] --> B["2. Identify 50–200 targets<br/>in one industry"]
B --> C["3. Reverse engineer<br/>thin claim charts"]
C --> D["4. Demand letter wave<br/>$50k–$500k each"]
D --> E{"5. Defendant<br/>response"}
E -->|"Pay quick<br/>~30%"| S["Settlement income"]
E -->|"Negotiate down<br/>~50%"| S
E -->|"Refuse<br/>~20%"| F["6. Litigation<br/>against holdouts"]
F --> G{"Cost picture<br/>clear"}
G -->|"Settle"| S
G -->|"Trial: rare"| T["Win or lose"]
T -->|"Win"| S
T -->|"Lose: patent invalidated"| K["Campaign dies"]
S --> R["7. Roll forward<br/>new targets, same patent"]
R --> B
style A fill:#fff7ed,stroke:#c2410c,color:#1c1917
style D fill:#fff7ed,stroke:#c2410c,color:#1c1917
style S fill:#d1fae5,stroke:#047857,color:#1c1917
style K fill:#fee2e2,stroke:#be123c,color:#1c1917
Read it as a funnel: 200 targets at the top, 50–80 settlements at the bottom, two or three actual lawsuits in the middle. The patent itself almost never gets tested at trial — by design. Trials risk invalidation, which kills the whole portfolio.
Stage 1: Buy a junk patent
The campaign starts with acquisition — before the troll has a target list. Sources, in rough order of volume:
- Bankruptcies. When tech companies fail, their IP gets liquidated. Nortel's $4.5B 2011 auction is the famous example. Patents that originally cost $50k–$200k to prosecute sell for a few thousand each in bulk lots after a bankruptcy.
- Individual inventors who can't afford to enforce themselves.
- Defensive divestments from large tech companies cleaning up their portfolios.
- Patent brokers like Ocean Tomo, ICAP, or specialized middlemen.
The acquisition cost is the defining number. A $30,000 patent that produces three settlements at $200,000 each is a 20× return — most of the work done by lawyers on contingency. The portfolio approach lets aggregators absorb the losers.
The selection criterion is breadth of reading, not patent quality. The troll wants a patent that can be argued to cover as many products as possible. Validity is a risk; breadth is yield.
Stage 2: Build a target list
Once the patent is in hand, the aggregator's team identifies an industry and a feature area where the patent reads. Common targets:
- E-commerce sites (matching, recommendation, comparison patents)
- SaaS tools (data-processing, workflow, analytics patents)
- Connected hardware (UI, wireless, sensor patents)
- Mid-market companies — large enough to settle, small enough to lack in-house IP teams
The list is 50–200 companies. Volume is the entire model. Even with a 5-figure average settlement against 30% of targets, a campaign generates seven figures of revenue.
Stage 3: Reverse engineer (thin claim charts)
Outside counsel and technical experts produce claim charts — spreadsheets that map each claim limitation to specific functionality in each accused product. The depth varies wildly:
- Quality campaigns — real engineers test products and build detailed mappings.
- Volume campaigns — thin charts that cite vague functional claims and generic screenshots from the defendant's website.
If you're staring at a demand letter, the depth of the claim chart is the strongest signal of which kind of campaign you're in. A defendant who pushes back hard sometimes gets the charts retracted because they're too thin to clear Rule 11 — that's a tell. Thin charts mean a settlement-shop, not a serious case.
Stage 4: The demand letter wave
The aggregator sends the same letter (with names swapped) to the entire target list within days of each other. Typical letter:
- States the patent number and a one-line invention summary
- Asserts the recipient's product infringes (often vaguely)
- References the patent's litigation history (intimidation)
- Quotes a "license fee" of $50,000–$500,000
- Includes a deadline 30–60 days out
Cost to the troll for the entire wave: a few hundred thousand dollars in attorney time, spread across 200 letters.
The asymmetry, in three numbers:
- Cost to send a demand letter: ~$2,000 per recipient
- Cost for you to defend a lawsuit through trial: $1M–$5M+
- Settlement demand: $50k–$500k — calibrated to land just below your defense cost
The number isn't picked randomly. It's picked to land in the rational-pay zone for a company of your size, given what defense would actually cost you.
Stage 5: The response funnel
Out of 200 letters, the response distribution is roughly this:
pie showData
title Demand-letter response distribution (typical)
"Pay quickly ($50k–$200k)" : 30
"Negotiate down ($20k–$80k)" : 50
"Ignore / refuse" : 20
Numbers vary by campaign but the shape is consistent:
- ~30% pay quickly. Often small companies that don't know what they're looking at, or larger companies whose risk-management math says "settle quickly is cheapest."
- ~50% negotiate the demand down by 50–70% and pay the lower number. The troll gladly accepts because their cost basis is so low.
- ~20% refuse outright — these become the litigation cohort.
The aggregator typically picks 10–30 from that 20% to actually sue, prioritizing companies with deep pockets and weak defenses. The rest either get a follow-up letter or disappear from the list.
Stage 6: The litigation pressure cooker
For defendants who are sued, the troll leverages the cost curve. Each phase of patent litigation has a characteristic cost spike:
flowchart LR
A["Filing → answer<br/>$25k–$100k"] --> B["Discovery<br/>$500k–$2M"]
B --> C["Markman / claim construction<br/>$200k–$500k"]
C --> D["Summary judgment<br/>$150k–$400k"]
D --> E["Trial<br/>$500k–$2M"]
style B fill:#fee2e2,stroke:#be123c,color:#1c1917
style E fill:#fee2e2,stroke:#be123c,color:#1c1917
The troll does NOT want to go to trial. Trial risks invalidation — and if the patent is invalidated, the entire portfolio (and the rest of the campaign) is dead. The troll wants the threat of trial to push settlement.
The settlement squeeze works like this:
The troll watches the defendant's discovery bills ramp. When defense costs hit $300k, the troll offers to settle for $250k. When bills hit $1M, the troll offers $400k. Each offer is just below current cumulative cost — psychologically, the defendant always feels like settling "saves" money.
This is why most patent suits settle during discovery, not at trial. The defendant has already spent the money — settlement is just acknowledging the spend and walking away. (See the patent litigation lifecycle for the phase-by-phase map.)
Stage 7: Roll forward
Once the campaign extracts its first batch of settlements, the aggregator:
- Pockets the revenue. Distributes to funders and the parent.
- Files against new entrants in the same industry. New companies built since the campaign started, or older companies that just hit the troll's radar.
- Starts a fresh campaign with the same patent, different target list — maybe a different industry vertical.
- Moves on to the next patent in the portfolio. Aggregators with hundreds of patents run dozens of overlapping campaigns at once.
Successful campaigns generate income for 3–7 years across dozens of defendants. By the time the patent expires, the troll has often multiplied the acquisition cost by 50× or more.
The math, in one composite example
A representative (illustrative, not a specific real case) campaign:
- Acquisition: 1 patent at $40,000.
- 150 demand letters sent. Total cost: ~$300,000 in attorney time.
- 50 settlements at an average of $80,000 each. Revenue: $4,000,000.
- 20 lawsuits filed against holdouts. 18 settle on average for $250,000 each. Revenue: $4,500,000.
- 2 lawsuits go to summary judgment. Both lose — patent invalidated. Cost: $1.5M each, $3,000,000 total. No revenue.
- Net to the aggregator + funders: roughly $5,000,000 over 3–4 years on a $40,000 acquisition.
Real campaigns vary wildly. Some patents generate $50M+ over their lifetime; others lose money. Aggregators absorb the losers because the wins pay for them several times over.
Why understanding this matters
Once you see the shakedown for what it is, your response framework changes:
| If you think the troll is… | You'll do… | Outcome |
|---|---|---|
| An unbeatable legal giant | Settle quickly, pay too much | Funds the next campaign against your peers |
| A serious patent enforcement claim | Hire counsel, fight on infringement, lose anyway | Spend $300k+, still settle |
| A volume-based extortion scheme | Refuse to negotiate against yourself, document everything, build leverage | Reset the troll's expected-value math; sometimes the case gets dropped |
You are not the only target. The troll's profitability depends on not having to actually litigate the patent. The troll wants you to settle. That asymmetry is also yours, if you use it.
A few tactics that exploit the cycle:
- Expose the campaign publicly. EFF's "Stupid Patent of the Month" series has killed multiple high-profile campaigns. Public attention raises the troll's litigation cost more than it raises yours.
- Coordinate with co-defendants. Joint defense groups split costs and share research. The troll's economic model depends on uncoordinated defendants making independent settle/fight decisions. (See What to do if you get a demand letter.)
- Threaten § 285 fees. A credible fee-shift threat changes the troll's expected-value math. (See § 285 fee-shifting.)
- File IPR. A $300k IPR can kill a patent the troll planned to monetize for years. (See Patent reviews and IPR estoppel for the trade-offs.)
- Pre-built invalidity dossiers. When § 102 / § 103 analyses exist before the troll asserts the patent, defense costs collapse. That's the structural fix this site is built for. Browse the database — if your patent is there, you start at $0 instead of $50,000 of prior-art research.
Bottom line
A patent-troll demand letter is the output of an industrial process — patent acquisition, target list, claim charts, demand wave, response funnel, settlement squeeze. It's not a serious legal threat tailored to you. It's volume arbitrage. Understanding the math is the first step to flipping the asymmetry, and the troll knows it — which is exactly why most campaigns rely on you not knowing.
If you're already in the cycle, read the demand-letter playbook. For the corporate-structure and litigation-finance side of how the operation actually runs, How patent trolls operate goes deeper on Stages 1–2.
The troll wants you to think you're up against a giant. You're up against a process. Processes can be broken.
This article is for general education and is not legal advice.